NOAA/WDS Paleoclimatology - LAKE STATUS RECORDS FROM THE FORMER SOVIET UNION AND MONGOLIA: DOCUMENTATION OF THE SECOND VERSION of the DATABASE

ABSTRACT SUPPLIED BY ORIGINATOR: This data set contains the information in Tarasov, P.E, et al, 1994 and 1996, "LAKE STATUS RECORDS FROM THE FORMER SOVIET UNION AND MONGOLIA: DATA BASE DOCUMENTATION", available from the World Data Center-A. The report documents reconstructions of long-term changes in lake status from 98 lakes from the former Soviet Union (FSU) and 5 lakes from Mongolia. Each lake basin is described in a separate file presented as Microsoft Word (*.doc) or text (*.txt) files. These text files contain the text of Tarosov, et al, available in published form from the World Data Center. The data base includes fully documented and coded records from 98 basins from the FSU and 5 basins from Mongolia. The data base was compiled by extracting information on changes in lake status from existing lithologic or biostratigraphic data. Most of the sites included in the data base have continuous Holocene records; many have lateglacial records and a few extend to beyond the last glacial maximum. The sites chosen for inclusion have been screened to meet standards of dating control and consistency among different climatic indicators. Records or parts of records where water depth appears to have been influenced by non-climatic factors, such a tectonism or human impact, or by factors where the climatic influence is indirect, such as sea-level changes and glacier fluctuations, have been excluded from the data base. A total of 23 basins have been excluded from the data base because they do not meet these standards. The major source of information on changes in lake status for most of the basins discussed in this report are changes in the nature of the bottom sediments, including lithology, grain size, organic content and chemical composition, as revealed in sediment cores. Palaeoecological evidence (aquatic pollen and macrofossils, diatoms, ostracodes, molluscs) have been used to provide corroborating evidence for the reconstructed changes in water depth. The reconstruction of changes in water depth at every site is based on the consensus interpretation of a minimum of two lines of evidence, following Harrison (1988), and Harrison and Digerfeldt (1993). For each site, an assessment of the relative water depth through time is made on the basis of all the available evidence. Where the evidence is limited or the range of variation is small, it may only be possible to distinguish two depth classes (deeper and shallower). There are sites, however, where it is possible to distinguish a greater number of depth-related differences in the geological and biostratigraphic data. This information is preserved by using a depth class categorisation that expands to the range demanded by the data. For each site, the lowest water depth recorded is coded as status class 1 and then successively deeper water phases are coded as 2, 3, 4 and so on until the maximum water depth recognised in the basin has been coded. It should be noted that it is rarely possible to quantify the changes in depth and the status classes do not represent a linear scale of depth changes. Assessments of lake status, using this flexible coding scheme, are made on a continuous basis. This information, along with the specific basis for the depth categorisation, is given in the documentation file for each basin. For comparison between sites, the lake status categories must be standardised. We use a 3 category scheme, where the status classes are defined as low (1), intermediate (2) and high (3). Various conventions can be used to define the boundaries of status classes. Here the boundaries were set so that for each lake record, the class "high" corresponded approximately to the upper quartile and "low" to the lower quartile of that lake's variation in level during the entire period of record, in order to ensure compatibility with the Oxford Lake-level Data Base (Street-Perrott et al., 1989). Status codings using this 3-category or "collapsed" coding scheme have been made at 500-yr intervals for each lake. These status codings are given in the data base and have been used for e.g. mapping purposes. The chronology of changes in lake status at individual sites is based on radiometric dating (radiocarbon, uranium/thorium), pollen correlation with a nearby radiocarbon-dated site, pollen correlation with a regional pollen chronostratigraphy, or a combination of these methods. These methods of dating a record are not all equally accurate, and therefore the basis for the chronology is indicated in the data base. The quality of the dating varies between different parts of the record from a single basin, and between the records from different basins. The quality of the dating control has been assessed at each 500-yr interval, using ranking schemes developed for the Cooperative Holocene Mapping (COHMAP) project (Webb, 1985). For data from continuous records, each interval was assigned a ranking (from 1 to 7) as follows: 1: Bracketing dates within a 2000-yr interval about the time being assessed 2: Bracketing dates, one within 2000 yr and the second within 4000 yr of the time being assessed 3: Bracketing dates within a 4000-yr interval about the time being assessed 4: Bracketing dates, one being within 4000 yr and the second being within 6000 yr of the time being assessed 5: Bracketing dates within a 6000-yr interval about the time being assessed 6: Bracketing dates, one within 6000 yr and the second within 8000 yr of the time being assessed 7: Poorly dated This scheme is only appropriate when sedimentation is continuous so that interpolation between radiometric (or biostratigraphic) dates is possible. At some sites, the evidence for changes in water depth was derived from discontinuous sources, such as shorelines, lake deposits above the modern lake and archaeological features. In some lake cores there is evidence of reworking, slumping, sedimentary hiatuses or marked variations in sedimentation rates, all of which make it difficult to interpolate between the dated intervals. Even when the sediment deposition can be assumed to be both continuous and uniform, there may be situations where there is a date very near one of the time intervals being coded but where the bracketing date is rather distant. Application of the scheme for continuous records would result in such an interval being inappropriately downweighted. For data from these types of discontinuous records, each 500-yr interval was assigned a ranking (from 1 to 7) as follows: 1: Date within 250 yr of the time being assessed 2: Date within 500 yr of the time being assessed 3: Date within 750 yr of the time being assessed 4: Date within 1000 yr of the time being assessed 5: Date within 1500 yr of the time being assessed 6: Date within 2000 yr of the time being assessed 7: Poorly dated.